When Viviane Soleil was an engineering student at Penn, she never imagined she would work with artificial intelligence. There was an AI course that didn’t seem very popular at the time, so she didn’t take it.
After a decades-long career, Sun is at the forefront of the AI revolution. Jabilan engineering, supply chain and manufacturing company where she serves as Senior Director of Data and AI, Enterprise Architecture and IT Transformation.
“I’ve managed big teams. I’ve managed big operations. But what I’ve always been interested in is managing change itself. Because we know the only thing that’s not going to change is change itself. That’s how I got started in AI,” she said.
Sun spoke to Wharton management professor Pierre Capelli for an episode of Where AI worksa podcast series produced by Wharton in collaboration with Accenture. (Listen to the episode.) She said Jabil first adopted AI five years ago, making it an early adopter. She remembers walking into a conference room with a group of colleagues to brainstorm ways to use emerging technology. The goal was not only to add value to their products, but also to improve the work process of their employees.
Jabil is a manufacturing company that makes products for its customers. Their first AI efforts focused on quality control, a critical issue in manufacturing. The first effort used machine learning to ensure that the colors they produced were exactly what their customers were asking for.
Sun cautioned that AI-assisted projects should be designed with long-term scalability and transformation in mind. This early victory led to wider use of AI. The company used computer vision, which trains AI to analyze visual images quickly and accurately, to assess visual defects in each product, a process that took about 20 seconds per item. The repetitive nature of the task makes workers prone to habituation, often resulting in inattention to critical details.
“After implementing the AI solution, we were able to remove some people from these repetitive and worthless (tasks). They were able to focus on the tasks they would like to do,” Sun said. “This has improved our inspection skills and saved our HR department time. Now they don’t need to search for as many inspectors.” But overall, jobs have not been lost.
Beyond getting the colors right, Jabil was able to replicate the tacit knowledge of experienced workers who could explain why colors weren’t right in different contexts. This project took time, but it highlighted how AI can take over certain tasks that relied on human experience developed over time.
“We use the data to analyze weather information so we can determine or suggest the correct amount of solutions to guide our workers, so we can get the exact color,” Sun said of refining the colors of certain products.
“We’re trying to use technology to solve a problem. We’re not trying to create a use case because we want to use technology.” — Viviane Soleil
Sun shared other ways Jabil, which has 140,000 employees worldwide, is implementing AI. Here are some key points from the conversation:
Start small, think big
Sun emphasized the importance of starting with small, manageable AI projects that demonstrate clear business value. She advised companies to anchor their AI initiatives in solving real-world business problems rather than following trends.
“I think the No. 1 suggestion is to start from business value rather than technology,” she said. “We’re trying to use technology to solve a problem. We’re not trying to create a use case because we want to use technology.”
AI agents become digital colleagues
Sun envisions an “explosion of AI agents” who will be treated the same as human employees, undergoing training and validation to align with company policies. These digital colleagues will take on repetitive tasks, make decisions and interact with internal and external stakeholders. This shift represents a major transformation in the way organizations view and manage their people.
“I believe this is going to happen because if you look at how agentic AI is transforming the world, it’s going to start to take control of decision-making,” she said. “It will be everywhere, in every sector, in our work, in our lives.”
Generative AI adds a layer of precision and validation
A very different use of AI at Jabil was to address rapid changes in tariffs and other international trade rules across their businesses. To improve the accuracy of international trade code assignment, Jabil combined machine learning with generative AI. Generative AI acts as a second opinion: first, machine learning estimates which code applies to a particular production, then this estimate is validated by querying documents via a chatbot. This layered approach improves reliability and reduces errors in critical business processes.
“If we get the same answer from machine learning and the chatbot, it’s more likely to be correct,” Sun said.
AI enhances human intelligence, not replaces it
Despite AI’s capabilities, Sun said its role is to augment human decision-making, not eliminate it. Especially in complex or variable scenarios, human judgment remains essential. This balanced perspective helps alleviate fears of layoffs and encourages human-AI collaborative workflows.
“AI is really about making people smarter. It helps people, but it can’t replace them in many ways today,” she said.
This article was partially AI-generated and edited (with additional writing) by Knowledge at Wharton staff. Read our AI policy here.